Re: [lwip-users] igmp status? versions?

[Jim Gibbons]

First and foremost, thank you for all the time you have been spending on lwip.  This must be quite a burden for you.  I was actually a little surprised to get a response from you directly.

I'm hoping that the IGMPv2 by Steve Reynolds will do the job for us.  If I'm right, then it will probably be a good thing for many other lwip users.

Here's the background on our situation:

We have developed a small appliance that sits on an ethernet for one of our customers.  It will normally acquire its address via DHCP.  Our customer then uses a host application to find all such appliances on the network and control them.  The message that is sent out to discover these clients is UDP/multicast.  This works fine so long as things are well localized on a single ethernet, but breaks down in a large enterprise environment.  We're hoping that by having our appliance enter the group for a specific multicast destination, we will be able to inform upstream routers of our need to see such messages, and thus get them from anywhere in a large enterprise.

As you can see, in this limited application, source filtering really isn't an issue, so v2 is all we need.  I suspect that this discovery problem is fairly common among small appliances, and thus fairly common among those using lwip in embedded systems.

Leon Woestenberg wrote:

Hello Jim,

Jim Gibbons wrote:

There is an open bug on IGMP integration, with a note indicating that the code contributed by Steve Reynolds would be used.  Does anyone know the status of this?

I am unfamiliar with typical practices in the use of IGMP.  Will v2 only be good enough?  That is, will it be OK to exclude v1 and v3?

Yes and no. I was planning on integrating Steve's implementation, then
refactor the code so that v3 support can be added.

v2 allows to join multicast groups (defined by their destination address)

v3 additionally allows to join source-specific multicasting (based
on source and destination).

I think v3 adds positive and negative filtering, which is the most
intrusive change.

Because we are not planning to use IGMP anytime soon, the status on Savannah is obsolote on this issue; I do not have the time to start work on this.

Regards,

Leon.

--

Jim Gibbons	address@hidden
Gibbons and Associates, Inc.	TEL: (408) 984-1441
900 Lafayette, Suite 704, Santa Clara, CA	FAX: (408) 247-6395